Updating the Featherweight Magnetic Switch

[Adrian A]

I have almost sold out my most recent production run of magnetic switches, and I'm refreshing the design with new parts, which should provide the following improvements:

• Lower on-resistance (< 7 mOhms)
• Higher allowable battery voltage (20V)
• Higher maximum current (36A)
• Lower standby power (~16 micro-Amps vs ~42 micro-Amps)
• Better clearance around mounting holes

The sensor and the behavior will the same. Pass either pole of a magnet by the sensor to turn it on, and pass it by again to turn it off. The mounting holes are still 4-40 with wire holes for optional soldering.

The new version should have about 1/3 the standby drain when connected to the battery (~16 micro-Amps vs ~42 micro-Amps), which means that if it is connected to a tiny 160mAhr battery it would take 47 days to discharge 10%, rather than 15 days.

Are there any other things that people wished were different about their magnetic switches? (mounting hole labels, etc)


Here are views of the bottom and top sides of the updated design, respectively:

 

[mpitfield]

Looks like the mounting holes are in alignment with each other on this one. Besides that is the footprint the same? Being and existing user with some low amp LiPos, I like the improvements, in particular the drain and lower on resistance. When will they be available?

 

[Worsaer]

Adrian, it would be nice if the default state when power is applied would be 'off' rather than 'on' (or if it could be an option).

 

[Adrian A]

Looks like the mounting holes are in alignment with each other on this one. Besides that is the footprint the same? Being and existing user with some low amp LiPos, I like the improvements, in particular the drain and lower on resistance. When will they be available?

Yes, the footprint is the same. It should be available in a few weeks.

 

[Adrian A]

Adrian, it would be nice if the default state when power is applied would be 'off' rather than 'on' (or if it could be an option).

I don't think it's possible without growing the size or the standby power consumption, but I'll take another look at it.

Also consider that there is a safety downside to having the default state be off, namely that if someone has a brownout due to charges firing or a temporary battery power disconnection in flight (not all that uncommon), having the switch turn on when power is re-applied could prevent a lawn dart. For ground safety, the best practice for this switch is to connect the battery and then turn the switch off before connecting charges. That way the altimeter need never be on with charges connected until you get to the pad. The switch can stay off for weeks while connected to the battery, so that provides a lot of flexibility in how and when the av-bay preparation is done.

-Adrian

 

[mpitfield]

Adrian can you do an overlay of the two rev boards to show the differences in mounting holes including any offset differences in relation to each other?

 

[Worsaer]

Really good points. Thanks Adrian.

I don't think it's possible without growing the size or the standby power consumption, but I'll take another look at it.

Also consider that there is a safety downside to having the default state be off, namely that if someone has a brownout due to charges firing or a temporary battery power disconnection in flight (not all that uncommon), having the switch turn on when power is re-applied could prevent a lawn dart. For ground safety, the best practice for this switch is to connect the battery and then turn the switch off before connecting charges. That way the altimeter need never be on with charges connected until you get to the pad. The switch can stay off for weeks while connected to the battery, so that provides a lot of flexibility in how and when the av-bay preparation is done.

-Adrian

 

[Adrian A]

Where did you get your avatar? That's very cool.

 

[Ravenex]

I have used your magnetic switches in all of my bays and plan to continue. My biggest issue is I want to solder my wires and screw it down, which make the screw not sit flat. I have taken to looping the wire and soldering it so it forms a washer of sorts... It would be nice if there was room to solder the wire flat on the board (not through) that would be out of the way of the screw. So maybe a rectangular pad outside the area of the screw head.

 

[Worsaer]

Where did you get your avatar? That's very cool.

Mine?

It's a T-Shirt found on the SF Embarcadero.

 

[OverTheTop]

Adrian, it would be nice if the default state when power is applied would be 'off' rather than 'on' (or if it could be an option).

I would vote for default being on for the same reason as someone else earlier in the thread.

 

[Adrian A]

Adrian can you do an overlay of the two rev boards to show the differences in mounting holes including any offset differences in relation to each other?

I did an overlay but compared to the most recent (and most common) version the outline and holes are completely identical

 

[Adrian A]

...It would be nice if there was room to solder the wire flat on the board (not through) that would be out of the way of the screw. So maybe a rectangular pad outside the area of the screw head.

I could keep solder mask off of the 3 corners. I have had concerns about the tight clearance between the bottom 2 pads so I don't think I want to add any more exposed conductors in that area, but the corners would be safe, if a bit small.

 

[KBlack]

Indicator LED visible from the activation edge. Placed where the existing G is would work. Then I can view the LED from a tiny hole in the side wall. (filled with transparent silicone if needed)

 

[Adrian A]

Indicator LED visible from the activation edge. Placed where the existing G is would work. Then I can view the LED from a tiny hole in the side wall. (filled with transparent silicone if needed)

Have you tried looking for it from a hole that's more in front of the mag switch rather than at the edge?

 

[KBlack]

I've done that... but you were taking suggestions. A perpendicular hole is much easier to drill and align to the tiny LED.

 

[BLKKROW]

Will the price remain the same, increase, decrease?

 

[mpitfield]

I have used your magnetic switches in all of my bays and plan to continue. My biggest issue is I want to solder my wires and screw it down, which make the screw not sit flat. I have taken to looping the wire and soldering it so it forms a washer of sorts... It would be nice if there was room to solder the wire flat on the board (not through) that would be out of the way of the screw. So maybe a rectangular pad outside the area of the screw head.

Hi Brian,

I hear your concern.

I sourced some ring terminals that fit the gauge wire I like to use, 20 gauge, and the ring part is almost the identical diameter as the round pads pads on the PCB including the center hole. I solder the ring terminals to the underside of the PCB so the are completely flush with the PCB, then crimp my wires on and add some colour matched heat shrink. When I mount it on the sled, as an extra step it is sandwiched between the standoff and the screw. Granted all of my standoffs and screws are nylon, so this doesn't provide much additional security but it does some.

Here is a pic of one in place. It is hard to see the terminals as they are on the bottom of the PCB...I have to remember to take a pic next time I make a set up. One other thing is that I always mock up the wires leading to the PCB so when I solder the ring terminals onto the pads they are pointed in the correct direction.

 

[Adrian A]

Will the price remain the same, increase, decrease?

I'm expecting to keep it the same, but I won't know for sure until I get an assembly quote.

 

[Adrian A]

I've done that... but you were taking suggestions. A perpendicular hole is much easier to drill and align to the tiny LED.

The LEDs shine much more brightly straight out from the board than they do to the side, so I'm not sure that trying to look at it edge-on would improve the visibility, even if it's on the edge of the circuit board. There are right-angle LEDs, but they would just shine against the wall for most people's av-bays. Many users (myself included) use the magnetic switch so that they don't have to drill and carefully align arming switch holes. I think the main use for the LED is to see that the switch is on when testing and troubleshooting the av-bay, and most people use an altimeter's beeps to verify that the rocket is ready to launch when at the pad. I know there are a lot of users with hearing loss, though, who use LEDs on an altimeter to verify that the rocket is ready. Does your altimeter have LEDs you can see at the pad? As a flier I would be uncomfortable launching with an altimeter that didn't give me any indication it's ready, even if I could verify that the power switch was turned on, because there are so many ways that its connections or charges could have problems.

Your suggestion may be valuable, but I think I would need to know more about your use-case to understand the advantage.

 

[Banzai88]

I have used your magnetic switches in all of my bays and plan to continue. My biggest issue is I want to solder my wires and screw it down, which make the screw not sit flat. I have taken to looping the wire and soldering it so it forms a washer of sorts... It would be nice if there was room to solder the wire flat on the board (not through) that would be out of the way of the screw. So maybe a rectangular pad outside the area of the screw head.

This.

 

[Ravenex]

I will confused ring terminals, though in some of my bays the extra length of stiff connector might be an issue.

 

[woferry]

Hi Adrian, I love the existing magnetic switch (have at least a dozen of them installed in different rockets), and all of these changes sound like welcome improvements, thanks! The drawing does seem to suggest that the holes are better-aligned, on my switches the 3 holes do not form a right triangle (or at least not parallel to the board edges), but it seems you're saying the new board's hole positions are the same as the old one? Same hole positions would certainly be ideal for my setups.

About the only comments I'd have are:

1) Could the small holes for the wire soldering be moved away from the actual rings for the screws? I.e. moved closer to the corners.

2) Could the plating extend all the way to the PCB corners on one or both sides of the PCB, ideally with a soldermask wall so that one could solder wires to these edges without the solder wanting to wick into the screw regions.

These two would seem to address some of the others concerns about soldering wires in the screw areas as it would provide a place to solder outside of that area, and I've attached a picture below showing a combined solution for #1 and #2 if it wasn't clear (lighter green is the exposed copper I added for soldering heavier gauge wire that might not fit in the small hole).

3) Could you add a second GND pad (just a flat pad, no screw hole) on the non-component side (a.k.a. bottom) in the 4th corner? And perhaps a 4th small hole for wire soldering in this corner, maybe just turn whichever of the two VIAs for the larger cap's GND in this corner into the larger plated hole, and perhaps put "G" silk next to it? In most cases I wire GND through your switch, so I'd like to have separate soldering points for the 'in' and 'out' GND connections. Perhaps like my #1 and #2 above the pad could be extended all the way to the PCB corner, though on the cap side I'd think you'd still want a soldermask wall to ensure the cap still gets a good fillet.

4) Can you see that Mike @ BAR gets a good supply? I'd like to be able to pick up parts like this on-site, and he used to carry these in his trailer which is how I first learned of them, but he's remarked that he's had difficulty getting in touch with you recently and hasn't been able to get stock in a while.

Personally I like the LED just as it is, I always line up my magnetic switches LED with the vent hole in my AvBays, which both gives me a visual clue of where to swipe the magnet and lets me see the LED through the hole as well for a quick visual check.

Oh, and curious, what is the existing switch resistance? Only potential issue I see here is that if it's a big difference it could also result in the altimeter seeing a higher-current spike during a deployment charge, so I'm wondering how much the delta is to assess the increased risk of damage to my altimeters. I hadn't considered before that the magnetic switch's resistance might be helping, the only altimeter I've blown during firing was one where I didn't use a mag switch, so I might have been getting a positive side-effect of its resistance, or if it was still pretty small maybe not.

And as far as ring terminals go, the small ones Dog House used to sell (should be part of Binder's lineup now) work perfectly for this application. :wink:

Thanks!!

 

[mpitfield]

I did an overlay but compared to the most recent (and most common) version the outline and holes are completely identical

The drawing does seem to suggest that the holes are better-aligned, on my switches the 3 holes do not form a right triangle (or at least not parallel to the board edges), but it seems you're saying the new board's hole positions are the same as the old one? Same hole positions would certainly be ideal for my setups.

No to beat a dead horse but my Mag switches sound like the same ones as woferry and I have about a dozen as well. I did the following drawing a while back as a template for drilling the holes.

 

[woferry]

No to beat a dead horse but my Mag switches sound like the same ones as woferry and I have about a dozen as well. I did the following drawing a while back as a template for drilling the holes.

View attachment 322232

Continuing with the beating ( ), interesting that you found the two 'bottom' holes (from your picture) to be parallel with the board edge, I didn't find that. The model I use for 3D printing sleds with holes for my switches worked best with these values:

magnetic_switch_hole_dims = [   // Distances (mm) from lower-left PCB corner
    [3, 4],                     // Lower-Left  (+In)
    [15, 10.8],                 // Upper-Right (Gnd)
    [14.8, 3],                  // Lower-Right (Out)
];

I called "lower left" the corner closest to the +In pin (your top-left), so my view is rotated 90° CCW from the way you pictured yours, meaning my +X axis is your vertical down, and +Y your right.

Regardless, it would be nice if you could provide the actual center dimensions, Adrian.

 

[Adrian A]

Hi Adrian, I love the existing magnetic switch (have at least a dozen of them installed in different rockets), and all of these changes sound like welcome improvements, thanks! The drawing does seem to suggest that the holes are better-aligned, on my switches the 3 holes do not form a right triangle (or at least not parallel to the board edges), but it seems you're saying the new board's hole positions are the same as the old one? Same hole positions would certainly be ideal for my setups.

About the only comments I'd have are:

1) Could the small holes for the wire soldering be moved away from the actual rings for the screws? I.e. moved closer to the corners.

2) Could the plating extend all the way to the PCB corners on one or both sides of the PCB, ideally with a soldermask wall so that one could solder wires to these edges without the solder wanting to wick into the screw regions.

These two would seem to address some of the others concerns about soldering wires in the screw areas as it would provide a place to solder outside of that area, and I've attached a picture below showing a combined solution for #1 and #2 if it wasn't clear (lighter green is the exposed copper I added for soldering heavier gauge wire that might not fit in the small hole).

3) Could you add a second GND pad (just a flat pad, no screw hole) on the non-component side (a.k.a. bottom) in the 4th corner? And perhaps a 4th small hole for wire soldering in this corner, maybe just turn whichever of the two VIAs for the larger cap's GND in this corner into the larger plated hole, and perhaps put "G" silk next to it? In most cases I wire GND through your switch, so I'd like to have separate soldering points for the 'in' and 'out' GND connections. Perhaps like my #1 and #2 above the pad could be extended all the way to the PCB corner, though on the cap side I'd think you'd still want a soldermask wall to ensure the cap still gets a good fillet.

4) Can you see that Mike @ BAR gets a good supply? I'd like to be able to pick up parts like this on-site, and he used to carry these in his trailer which is how I first learned of them, but he's remarked that he's had difficulty getting in touch with you recently and hasn't been able to get stock in a while.

Personally I like the LED just as it is, I always line up my magnetic switches LED with the vent hole in my AvBays, which both gives me a visual clue of where to swipe the magnet and lets me see the LED through the hole as well for a quick visual check.

Oh, and curious, what is the existing switch resistance? Only potential issue I see here is that if it's a big difference it could also result in the altimeter seeing a higher-current spike during a deployment charge, so I'm wondering how much the delta is to assess the increased risk of damage to my altimeters. I hadn't considered before that the magnetic switch's resistance might be helping, the only altimeter I've blown during firing was one where I didn't use a mag switch, so I might have been getting a positive side-effect of its resistance, or if it was still pretty small maybe not.

And as far as ring terminals go, the small ones Dog House used to sell (should be part of Binder's lineup now) work perfectly for this application. :wink:

Thanks!!

Thanks for taking the time for all the good feedback.

1 & 2. I have already updated the design to make the whole corner exposed. Putting the wire through-hole in the corner too is a great idea I should have thought of before. I've always just soldered the wires and expoxied down my switches in place for my own rockets.

The trouble with moving the wire holes away from the mounting hole pads is that with the current wire hole size, mounting screw location and board outline, it doesn't quite all fit. The board fabrication houses require a minimum copper ring width around through-hole pads and clearance from the board edge, which makes it different from your picture. I could scrunch in the mounting holes inward a little, expand the board outline or leave the wire holes overlapping the mounting hole pad. I think moving the wire holes away from the mounting hole pads is worth doing, and I'm inclined to move the mounting holes inward rather than increase the board outline, on the theory that people using new switches for new av-bays will vastly outnumber people who would want to use new switches with their old mounting hole locations. But if there are a lot of people who would much prefer growing the board size a little I'll do that.

In case anyone is curious, the way I install mag switches in my own rockets is to solder to the mounting holes and use 5-minute epoxy to mount the board.

3. Another good idea. Done.

4. My dealer relationships work out when he or she is on-site for enough launches and has enough volume to make it worthwhile for both of us. I don't remember ever really getting started with BAR.

The previous version switch used a MOSFET that had an Rdson as low as 6 mOhms when used with a 9V input but 10-20 mOhms when driven between 3 and 4 volts. The new FET has a range of 4-6 mOhms for the whole input voltage range of 3.5V - 20V. If you're on the edge about over-current of your altimeter outputs, a smaller battery or skinnier wire will make your rocket safer. I prefer 26 gauge and counsel against anything larger than 22. Lipos are great but for a Raven altimeter they need to be 165 mAhrs or smaller.

 

[Adrian A]

No to beat a dead horse but my Mag switches sound like the same ones as woferry and I have about a dozen as well. I did the following drawing a while back as a template for drilling the holes.

Thanks for providing the template. There were a few versions of the older switches. I think one of them had funky hole locations that I subsequently fixed.

 

[Adrian A]

Here is the updated design. Bottom on the left, top on the right. The hatched areas are solder-coated copper.



The holes are 0.14" from each edge of the board, which is 0.54" x 0.72".

 

[woferry]

Looks good to me, thanks!

 

[Wayco]

I really like what you are doing here Adrian. Increasing the area and moving the hole outside the screw seat is a good improvement, and having another ground pad is too, but the best thing you have done is reducing the current draw in standby. Now I won't have to remove batteries between the monthly launches.
Keep us posted on when these become available, I have a couple more projects that need them.